{"id":"W4400485045","doi":"10.1145/3664646.3664772","title":"Chain of Targeted Verification Questions to Improve the Reliability of Code Generated by LLMs","year":2024,"lang":"en","type":"article","venue":"","topic":"Software Testing and Debugging Techniques","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Reliability (semiconductor); Computer science; Code (set theory); Chain (unit); Reliability engineering; Risk analysis (engineering); Programming language; Business; Engineering","routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.005400572,0.002393904,0.001396662,0.003126505,0.0006336074,0.001134181,0.002744624,0.001986438,0.005654948],"category_scores_gemma":[0.05278101,0.0007795744,0.001836605,0.0008775548,0.000975633,0.002808943,0.002971777,0.001789753,0.003049952],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009495663,"about_ca_system_score_gemma":0.002352406,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003077489,"about_ca_topic_score_gemma":0.005268756,"domain_scores_codex":[0.9913102,0.003263975,0.000681724,0.002211073,0.002189979,0.0003430201],"domain_scores_gemma":[0.9513153,0.03100107,0.004114128,0.006175181,0.00667866,0.0007155886],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001658953,0.0006639664,0.0168368,0.002475603,0.0001550991,0.001103026,0.002707157,0.01882732,0.06691499,0.003290207,0.02397619,0.8613907],"study_design_scores_gemma":[0.0005135204,0.001237486,0.01193837,0.0004482871,0.0002996393,0.001286612,0.0009565541,0.8142296,0.1106291,0.01670956,0.04156511,0.0001862358],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1100399,0.001304917,0.7816386,0.001087519,0.0002341192,0.00121942,0.001972915,0.09980861,0.002694003],"genre_scores_gemma":[0.3166335,0.0002646459,0.6647909,0.0008792985,0.0001275356,0.0006661804,0.00683635,0.004997216,0.004804309],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005654948,"threshold_uncertainty_score":0.02856129,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01257638812373305,"score_gpt":0.2681475290310201,"score_spread":0.255571140907287,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}